%load_ext autoreload
%autoreload 2

import pandas as pd
import numpy as np 
from matplotlib import pyplot as plt 
import seaborn as sbn 

import statsmodels.api as sm
from sklearn.preprocessing import OneHotEncoder

%aimport HER2_classifier

class myspace(object): 
    def __init__(self): 
        pass

args                    =       myspace()
args.data               =       ['./HER2_SKBR3_data_6-7-21/']
args.out                =       ['./output/'] 
args.drug               =       ['Neratinib']   # ['Trastuzumab']
args.sensitive_line     =       ['WT']
args.resistant_line     =       ['T798I']
args.load               =       ['normalized']  # ['raw']
args.nclus              =       [15]
args.resample_sz        =       [125]
args.burnin             =       [0]

Load Data¶

data, clover_sel, mscarl_sel = HER2_classifier.load_data(args)
print('len selector:', len(clover_sel))
print(clover_sel[0:5])
print(mscarl_sel[0:5])
data.head()

loading data...
# of datasets to load: 13
mapping drug names to one name...
len selector: 135
['0_x', '15_x', '30_x', '45_x', '60_x']
['0_y', '15_y', '30_y', '45_y', '60_y']

data.drug.unique()

array(['untreated', '10nm_neratinib', '10ug_ml_trastuzumab'], dtype=object)

data.cell_line.unique()

array(['SKBR3'], dtype=object)

data.mutant.unique()

array(['WT', 'T798I', 'ND611', 'ND648', 'S310Y', 'T306M', 'L313V',
       'T733I', 'L755T759DEL', 'I767M', 'A355QFS76', 'D639E', 'Y835F',
       'V842I', 'R678Q', 'V424I', 'D769Y', 'E717K', 'E744D', 'G621AFS31',
       'I435L', 'I886M', 'L43V', 'P416T', 'P702L', 'R203P', 'R340G',
       'R456C', 'R47C', 'S609C', 'T479M', 'W906C', 'Y64F', 'D277H',
       'D638H', 'D873G', 'G1055C', 'G778S779INSLPS', 'H470Q', 'H878Y',
       'L662V', 'M833I', 'L785F', 'L869R', 'P122L', 'S305C', 'W482GFS74',
       'E939G', 'G727A', 'M889I', 'N488S', 'Q1206K', 'W825', 'A1039T',
       'E975Q', 'S310F', 'R34W', 'R929WL', 'S963', 'V797A',
       'Y722V773INSRDGE', 'L1197F', 'R1153L', 'Y1127A1129DEL', 'R897G',
       'P523S', 'D769N', 'E770A771INSGIRD', 'N427K', 'P232S', 'R190Q',
       'A1160V', 'T297I', 'T862S', 'V697L', 'P579L', 'R288Q', 'A37T',
       'D933Y', 'E405D', 'F976L', 'L755S'], dtype=object)

Filter NA¶

data, clover_sel, mscarl_sel = HER2_classifier.filter_na(data, args, clover_sel, mscarl_sel)
print('len selector:', len(clover_sel))
print(clover_sel[0:5])
print(mscarl_sel[0:5])
data.head()

filtering to drug and removing NAs...
Data shape (untreated + drug): (23672, 398)
length of time series BEFORE removing time points with NA 135
length of time series AFTER removing time points with NA 135
len selector: 135
['0_x' '15_x' '30_x' '45_x' '60_x']
['0_y' '15_y' '30_y' '45_y' '60_y']

data.groupby(['mutant', 'drug']).count()['track_index'].sort_values().head(15)

mutant           drug          
E770A771INSGIRD  10nm_neratinib    17
W825             untreated         19
W482GFS74        untreated         19
T862S            untreated         20
P702L            untreated         21
M889I            untreated         21
                 10nm_neratinib    21
S963             untreated         22
P122L            untreated         23
E770A771INSGIRD  untreated         24
W825             10nm_neratinib    25
S305C            untreated         26
T862S            10nm_neratinib    29
R897G            untreated         34
W482GFS74        10nm_neratinib    37
Name: track_index, dtype: int64

data.groupby(['mutant', 'drug']).count()['track_index'].sort_values().tail(15)

mutant          drug          
S310Y           10nm_neratinib     195
A37T            10nm_neratinib     210
P579L           10nm_neratinib     219
L662V           10nm_neratinib     225
H470Q           10nm_neratinib     239
T479M           10nm_neratinib     256
R678Q           10nm_neratinib     274
G778S779INSLPS  10nm_neratinib     284
D277H           10nm_neratinib     309
ND611           untreated          810
                10nm_neratinib    1159
T798I           untreated         1194
                10nm_neratinib    1297
WT              untreated         1369
                10nm_neratinib    1800
Name: track_index, dtype: int64

plt.hist(data[lambda x: ~x.mutant.isin(['WT', 'T798I', 'ND611'])].groupby(['mutant', 'drug']).count()['track_index'])

(array([26., 33., 32., 27., 20., 11.,  3.,  2.,  2.,  2.]),
 array([ 17. ,  46.2,  75.4, 104.6, 133.8, 163. , 192.2, 221.4, 250.6,
        279.8, 309. ]),
 <BarContainer object of 10 artists>)

Add `Burn-in`¶

Remove the first few time points

clover_sel, mscarl_sel = HER2_classifier.burnin(args, clover_sel, mscarl_sel)
print('len selector:', len(clover_sel))
print('len selector:', len(mscarl_sel))
print(clover_sel[0:5])
print(mscarl_sel[0:5])

adding burnin of 0
len selector: 135
len selector: 135
['0_x' '15_x' '30_x' '45_x' '60_x']
['0_y' '15_y' '30_y' '45_y' '60_y']

Resample time-series¶

X_train = HER2_classifier.resample(data, args, clover_sel, mscarl_sel)

resampling time series...
Training data shape BEFORE resampling: (23672, 135, 2)
Training data shape AFTER resampling: (23672, 125, 2)

Fit the time-series K-means clustering¶

y_pred, km = HER2_classifier.fit_timeseries_kmeans(args, X_train, plot=True, save=None)

performing time-series kmeans clustering...

5.459 --> 3.927 --> 3.838 --> 3.803 --> 3.780 --> 3.762 --> 3.750 --> 3.743 --> 3.738 --> 3.734 --> 3.731 --> 3.729 --> 3.727 --> 3.725 --> 3.723 --> 3.721 --> 3.720 --> 3.718 --> 3.717 --> 3.716 --> 3.715 --> 3.714 --> 3.713 --> 3.713 --> 3.712 --> 3.712 --> 3.711 --> 3.711 --> 3.710 --> 3.710 --> 3.709 --> 3.709 --> 3.709 --> 3.708 --> 3.708 --> 3.707 --> 3.707 --> 3.706 --> 3.706 --> 3.706 --> 3.706 --> 3.706 --> 3.706 --> 3.705 --> 3.705 --> 3.705 --> 3.705 --> 3.705 --> 3.705 --> 3.705 -->

Calculate the cluster proportion¶

...within each experiment.

cm, lb = HER2_classifier.quantify_cluster_prop(args, data, y_pred)
cm.shape

quantifying experiment by cluster proportions...

(252, 15)

Visualize the cluster co-occurence within experiment¶

HER2_classifier.plot_cluster_corr(cm, save=None)

<Figure size 504x504 with 0 Axes>

Dimensionality Reduction¶

res, pca = HER2_classifier.reduce_dim(args, cm, lb, plot=True, save=None)

performing dim. reduction (pca)...
PCA explained variance ratio: [0.49678721 0.11723573]
PC shape: (252, 2)

pc_loadings = pd.DataFrame({'clus_feat': range(pca.components_.shape[1]), 'PC1':pca.components_[0], 'PC2':pca.components_[1]})
pc_loadings.head()

plt.figure(figsize=(10,7))
sbn.barplot(x='clus_feat', y='PC1', data=pc_loadings, order=pc_loadings.sort_values(by='PC1').clus_feat)
plt.show()

plt.figure(figsize=(10,7))
sbn.barplot(x='clus_feat', y='PC2', data=pc_loadings, order=pc_loadings.sort_values(by='PC2').clus_feat)
plt.show()

res

Check for Batch Effects¶

batch_res = HER2_classifier.check_batch_effects(args, res, plot=True, save=None)
batch_res.head()

['H210326_Set2' 'H210409_Set1' 'H210205_Set2_2' 'H210409_Set2'
 'H210423_Set1' 'H210219_Set1' 'H210401_Set2' 'H210212_Set1'
 'H210219_Set2' 'H210326_Set1' 'H210205_Set1_2' 'H210401_Set1'
 'H210212_Set2']
['H210205_Set1_2', 'H210205_Set2_2', 'H210212_Set1', 'H210212_Set2', 'H210219_Set1', 'H210219_Set2', 'H210326_Set1', 'H210326_Set2', 'H210401_Set1', 'H210401_Set2', 'H210409_Set1', 'H210409_Set2', 'H210423_Set1', 'neratinib', 'untreated', 'T798I', 'WT']
feature order:
	 x1  ->  H210205_Set1_2
	 x2  ->  H210205_Set2_2
	 x3  ->  H210212_Set1
	 x4  ->  H210212_Set2
	 x5  ->  H210219_Set1
	 x6  ->  H210219_Set2
	 x7  ->  H210326_Set1
	 x8  ->  H210326_Set2
	 x9  ->  H210401_Set1
	 x10  ->  H210401_Set2
	 x11  ->  H210409_Set1
	 x12  ->  H210409_Set2
	 x13  ->  H210423_Set1
	 x14  ->  neratinib
	 x15  ->  untreated
	 x16  ->  T798I
	 x17  ->  WT
---------------------------------------------------------------------------
PC1 ANOVA
---------------------------------------------------------------------------
                            OLS Regression Results                            
==============================================================================
Dep. Variable:                      y   R-squared:                       0.885
Model:                            OLS   Adj. R-squared:                  0.842
Method:                 Least Squares   F-statistic:                     20.43
Date:                Mon, 07 Jun 2021   Prob (F-statistic):           3.03e-13
Time:                        11:03:04   Log-Likelihood:                 73.208
No. Observations:                  52   AIC:                            -116.4
Df Residuals:                      37   BIC:                            -87.15
Df Model:                          14                                         
Covariance Type:            nonrobust                                         
==============================================================================
                 coef    std err          t      P>|t|      [0.025      0.975]
------------------------------------------------------------------------------
const         -0.0119      0.005     -2.538      0.015      -0.021      -0.002
x1             0.0320      0.034      0.948      0.349      -0.036       0.100
x2            -0.0432      0.034     -1.280      0.208      -0.111       0.025
x3             0.0031      0.034      0.092      0.928      -0.065       0.071
x4            -0.0023      0.034     -0.068      0.946      -0.071       0.066
x5             0.0417      0.034      1.236      0.224      -0.027       0.110
x6            -0.0022      0.034     -0.065      0.949      -0.070       0.066
x7             0.0128      0.034      0.380      0.706      -0.056       0.081
x8            -0.0278      0.034     -0.825      0.415      -0.096       0.041
x9            -0.0042      0.034     -0.124      0.902      -0.072       0.064
x10           -0.0248      0.034     -0.735      0.467      -0.093       0.044
x11            0.0037      0.034      0.109      0.914      -0.065       0.072
x12           -0.0211      0.034     -0.626      0.535      -0.089       0.047
x13            0.0204      0.034      0.606      0.548      -0.048       0.089
x14            0.1520      0.010     15.179      0.000       0.132       0.172
x15           -0.1638      0.010    -16.367      0.000      -0.184      -0.144
x16           -0.0462      0.010     -4.610      0.000      -0.066      -0.026
x17            0.0343      0.010      3.422      0.002       0.014       0.055
==============================================================================
Omnibus:                       25.902   Durbin-Watson:                   0.697
Prob(Omnibus):                  0.000   Jarque-Bera (JB):                4.150
Skew:                          -0.070   Prob(JB):                        0.126
Kurtosis:                       1.623   Cond. No.                     4.76e+16
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The smallest eigenvalue is 4.76e-32. This might indicate that there are
strong multicollinearity problems or that the design matrix is singular.
---------------------------------------------------------------------------
PC2 ANOVA
---------------------------------------------------------------------------
                            OLS Regression Results                            
==============================================================================
Dep. Variable:                      y   R-squared:                       0.714
Model:                            OLS   Adj. R-squared:                  0.606
Method:                 Least Squares   F-statistic:                     6.608
Date:                Mon, 07 Jun 2021   Prob (F-statistic):           2.00e-06
Time:                        11:03:04   Log-Likelihood:                 86.067
No. Observations:                  52   AIC:                            -142.1
Df Residuals:                      37   BIC:                            -112.9
Df Model:                          14                                         
Covariance Type:            nonrobust                                         
==============================================================================
                 coef    std err          t      P>|t|      [0.025      0.975]
------------------------------------------------------------------------------
const          0.0323      0.004      8.823      0.000       0.025       0.040
x1             0.0370      0.026      1.404      0.169      -0.016       0.090
x2            -0.0035      0.026     -0.131      0.896      -0.057       0.050
x3            -0.0093      0.026     -0.353      0.726      -0.063       0.044
x4            -0.0078      0.026     -0.295      0.770      -0.061       0.046
x5             0.0182      0.026      0.692      0.494      -0.035       0.072
x6            -0.0144      0.026     -0.547      0.588      -0.068       0.039
x7             0.0185      0.026      0.702      0.487      -0.035       0.072
x8            -0.0129      0.026     -0.489      0.628      -0.066       0.040
x9            -0.0201      0.026     -0.762      0.451      -0.073       0.033
x10            0.0284      0.026      1.078      0.288      -0.025       0.082
x11            0.0203      0.026      0.772      0.445      -0.033       0.074
x12           -0.0037      0.026     -0.139      0.890      -0.057       0.050
x13           -0.0186      0.026     -0.705      0.485      -0.072       0.035
x14            0.0155      0.008      1.983      0.055      -0.000       0.031
x15            0.0168      0.008      2.147      0.038       0.001       0.033
x16            0.0869      0.008     11.110      0.000       0.071       0.103
x17           -0.0546      0.008     -6.980      0.000      -0.070      -0.039
==============================================================================
Omnibus:                        1.596   Durbin-Watson:                   1.874
Prob(Omnibus):                  0.450   Jarque-Bera (JB):                1.320
Skew:                          -0.387   Prob(JB):                        0.517
Kurtosis:                       2.907   Cond. No.                     4.76e+16
==============================================================================

Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The smallest eigenvalue is 4.76e-32. This might indicate that there are
strong multicollinearity problems or that the design matrix is singular.

Train classifier on `[positive/negative]` controls¶

model, accuracy = HER2_classifier.train_classifier(args, res, plot=True, save=None)

training classifier...
sensitive line: 	 WT
resistant line: 	 T798I
drug:		 neratinib
drug + WT df size:  (26, 5)
X train shape: (26, 2)
# neg class (resistant): 13
# pos class (sensitive): 13

Assign mutant sensitivity/resistance calls¶

prob_res = HER2_classifier.predict_mutants(args, model, res, batch_res)
prob_res

predicting unlabeled sensitivities...

prob_res.tail(10)

plt.figure(figsize=(10,10))
sbn.scatterplot(x='pc1', y='pc2', data=prob_res, hue='prob_res', style='call', s=300)
plt.show()

0.6/0.4

1.4999999999999998

np.log2(0.4/0.6)

-0.5849625007211561

	track_index	0_x	15_x	30_x	45_x	60_x	75_x	90_x	105_x	120_x	...	2805_y	2820_y	2835_y	2850_y	2865_y	2880_y	2895_y	drug	cell_line	mutant
0	3_1000011860_skbr3_erk_akt__WT_untreated	0.836900	0.838897	0.557456	0.852664	1.137005	0.798656	0.696818	0.596282	0.495039	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
1	79_1000288819_skbr3_erk_akt__WT_untreated	0.615120	0.775863	0.840381	0.828898	0.797149	0.827444	0.779779	0.848228	0.868804	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
2	78_1000285319_skbr3_erk_akt__WT_untreated	0.790633	0.865733	0.793790	0.890870	0.870234	0.917848	0.700682	1.004571	0.972406	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
3	78_1000282488_skbr3_erk_akt__WT_untreated	0.919838	0.836807	0.853458	0.712794	0.662686	0.630794	0.635830	0.723558	0.749466	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
4	1_1000002701_skbr3_erk_akt__WT_untreated	0.822612	0.834595	0.920568	0.968030	0.896872	0.906976	0.933929	0.925297	0.913278	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT

	track_index	0_x	15_x	30_x	45_x	60_x	75_x	90_x	105_x	120_x	...	2805_y	2820_y	2835_y	2850_y	2865_y	2880_y	2895_y	drug	cell_line	mutant
0	3_1000011860_skbr3_erk_akt__WT_untreated	0.836900	0.838897	0.557456	0.852664	1.137005	0.798656	0.696818	0.596282	0.495039	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
1	79_1000288819_skbr3_erk_akt__WT_untreated	0.615120	0.775863	0.840381	0.828898	0.797149	0.827444	0.779779	0.848228	0.868804	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
2	78_1000285319_skbr3_erk_akt__WT_untreated	0.790633	0.865733	0.793790	0.890870	0.870234	0.917848	0.700682	1.004571	0.972406	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
3	78_1000282488_skbr3_erk_akt__WT_untreated	0.919838	0.836807	0.853458	0.712794	0.662686	0.630794	0.635830	0.723558	0.749466	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT
4	1_1000002701_skbr3_erk_akt__WT_untreated	0.822612	0.834595	0.920568	0.968030	0.896872	0.906976	0.933929	0.925297	0.913278	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	untreated	SKBR3	WT

	clus_feat	PC1	PC2
0	0	-0.264865	-0.152827
1	1	-0.064274	0.208965
2	2	-0.274968	-0.140808
3	3	0.320203	0.032544
4	4	-0.141994	-0.269706

	constant	H210205_Set1_2	H210205_Set2_2	H210212_Set1	H210212_Set2	H210219_Set1	H210219_Set2	H210326_Set1	H210326_Set2	H210401_Set1	H210401_Set2	H210409_Set1	H210409_Set2	H210423_Set1	neratinib	untreated	T798I	WT	PC
0	1.548786e-02	0.349278	0.208416	0.927536	0.945985	0.224213	0.948919	0.706340	0.414765	0.902293	0.467051	0.913812	0.534893	0.548480	1.798289e-17	1.573839e-18	4.674795e-05	1.532178e-03	1
1	1.242664e-10	0.168790	0.896442	0.726417	0.769692	0.493509	0.587571	0.486851	0.627625	0.451049	0.288229	0.445069	0.889915	0.485351	5.483875e-02	3.841371e-02	2.411433e-13	3.012377e-08	2

	pc1	pc2	treatment	cell_line	exp_set	prob_res	prob_sens	call	odds_ratio	batch	PC1_batch_pval	PC1_batch_flag	PC2_batch_pval	PC2_batch_flag
64	0.232905	-0.027193	neratinib	P579L	H210423_Set1	0.044584	0.955416	sens	0.046664	H210423_Set1	0.548480	False	0.485351	False
85	0.217037	-0.027397	neratinib	T306M	H210205_Set1_2	0.044895	0.955105	sens	0.047005	H210205_Set1_2	0.349278	False	0.168790	False
40	0.237333	-0.021065	neratinib	M833I	H210219_Set2	0.045076	0.954924	sens	0.047204	H210219_Set2	0.948919	False	0.587571	False
13	0.216064	-0.032084	neratinib	E405D	H210409_Set2	0.045154	0.954846	sens	0.047290	H210409_Set2	0.534893	False	0.889915	False
72	0.242047	-0.008056	neratinib	R340G	H210212_Set2	0.046946	0.953054	sens	0.049259	H210212_Set2	0.945985	False	0.769692	False
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
76	0.099087	0.061844	neratinib	R678Q	H210205_Set2_2	0.634225	0.365775	res	1.733922	H210205_Set2_2	0.208416	False	0.896442	False
91	0.096304	0.072438	neratinib	V797A	H210401_Set1	0.707492	0.292508	res	2.418710	H210401_Set1	0.902293	False	0.451049	False
92	0.061079	0.042302	neratinib	V842I	H210205_Set2_2	0.713117	0.286883	res	2.485736	H210205_Set2_2	0.208416	False	0.896442	False
25	0.091012	0.072935	neratinib	G778S779INSLPS	H210219_Set2	0.735133	0.264867	res	2.775485	H210219_Set2	0.948919	False	0.587571	False
38	0.082803	0.141710	neratinib	L785F	H210326_Set1	0.920393	0.079607	res	11.561642	H210326_Set1	0.706340	False	0.486851	False

	pc1	pc2	treatment	cell_line	exp_set
0	0.161031	0.048727	neratinib	A1039T	H210326_Set2
1	-0.223676	-0.067015	untreated	A1039T	H210326_Set2
2	0.246843	-0.030411	neratinib	A1160V	H210409_Set1
3	-0.236713	-0.063892	untreated	A1160V	H210409_Set1
4	0.163846	0.052475	neratinib	A355QFS76	H210205_Set2_2
...	...	...	...	...	...
247	-0.170035	0.003534	untreated	Y64F	H210219_Set1
248	0.102057	0.055529	neratinib	Y722V773INSRDGE	H210401_Set1
249	-0.103252	0.083981	untreated	Y722V773INSRDGE	H210401_Set1
250	0.147673	0.099242	neratinib	Y835F	H210205_Set2_2
251	-0.170091	0.112895	untreated	Y835F	H210205_Set2_2

	pc1	pc2	treatment	cell_line	exp_set	prob_res	prob_sens	call	odds_ratio	batch	PC1_batch_pval	PC1_batch_flag	PC2_batch_pval	PC2_batch_flag
35	0.147088	0.076456	neratinib	L755S	H210409_Set2	0.397996	0.602004	sens	0.661119	H210409_Set2	0.534893	False	0.889915	False
37	-0.189156	-0.057121	neratinib	L755T759DEL	H210205_Set1_2	0.466859	0.533141	sens	0.875677	H210205_Set1_2	0.349278	False	0.168790	False
99	0.147673	0.099242	neratinib	Y835F	H210205_Set2_2	0.534418	0.465582	res	1.147850	H210205_Set2_2	0.208416	False	0.896442	False
20	0.126045	0.074688	neratinib	F976L	H210409_Set2	0.535775	0.464225	res	1.154130	H210409_Set2	0.534893	False	0.889915	False
98	0.102057	0.055529	neratinib	Y722V773INSRDGE	H210401_Set1	0.581248	0.418752	res	1.388048	H210401_Set1	0.902293	False	0.451049	False
76	0.099087	0.061844	neratinib	R678Q	H210205_Set2_2	0.634225	0.365775	res	1.733922	H210205_Set2_2	0.208416	False	0.896442	False
91	0.096304	0.072438	neratinib	V797A	H210401_Set1	0.707492	0.292508	res	2.418710	H210401_Set1	0.902293	False	0.451049	False
92	0.061079	0.042302	neratinib	V842I	H210205_Set2_2	0.713117	0.286883	res	2.485736	H210205_Set2_2	0.208416	False	0.896442	False
25	0.091012	0.072935	neratinib	G778S779INSLPS	H210219_Set2	0.735133	0.264867	res	2.775485	H210219_Set2	0.948919	False	0.587571	False
38	0.082803	0.141710	neratinib	L785F	H210326_Set1	0.920393	0.079607	res	11.561642	H210326_Set1	0.706340	False	0.486851	False

Load Data¶

Filter NA¶

Add Burn-in¶

Resample time-series¶

Fit the time-series K-means clustering¶

Calculate the cluster proportion¶

Visualize the cluster co-occurence within experiment¶

Dimensionality Reduction¶

Check for Batch Effects¶

Train classifier on [positive/negative] controls¶

Assign mutant sensitivity/resistance calls¶

Add `Burn-in`¶

Train classifier on `[positive/negative]` controls¶